[felicemastronzo]

Hi all,

I have a gearbox filled with oil. The motion of the internal gears heat the oil. I want to calculate the temperature distribution in the oil bath. Which BC should I set on the external boundaries (the case of the gearbox)? I do not know the temperature of the walls but only the temperature of the ambient (air). I know that the heat transfer coeff. h is depending on the delta T between the walls and the external air. If I want define that coefficient according to a known law, what should I do? Can someone help me writing an UDF?

Regards
F

air | wall | oil | gears

[Zigainer]

Hi,

I haven't done conjugated heat transfer with fluent, but I did it with CFX .... so I might be able help you.

First question is, what are you interested about? Are you interested in a temperature distribution after a specific time? -> Transient run or "just" in a steady state case?

As boundary condition: for the wall | air interaction use a heat flux. This heat flux is, as you mentioned, not constant it is rather dependent on the surface temperature. So you have to create a "heat-flux-function"(Temperature) .... I haven't done this in fluent, but in CFX this is quite simple. Hope some other guys here can help you out with a UDF.

And how do you define your "heat-source"? Are you simulating the gears as well, I guess the gears would lead to some motion of the nearby oil ....

[felicemastronzo]

Originally Posted by Zigainer

Hi,

I haven't done conjugated heat transfer with fluent, but I did it with CFX .... so I might be able help you.

First question is, what are you interested about? Are you interested in a temperature distribution after a specific time? -> Transient run or "just" in a steady state case?

As boundary condition: for the wall | air interaction use a heat flux. This heat flux is, as you mentioned, not constant it is rather dependent on the surface temperature. So you have to create a "heat-flux-function"(Temperature) .... I haven't done this in fluent, but in CFX this is quite simple. Hope some other guys here can help you out with a UDF.

And how do you define your "heat-source"? Are you simulating the gears as well, I guess the gears would lead to some motion of the nearby oil ....

Hi, thans for the answer.
In my case I have a single gear that is rotating in the oil bath, so the heat is directly generated by the motion of the gear. I have already performed transient simulations without the energy equations, so I already know the power losses (--> heat source). In order to calculate them a transient simulation was mandatory, but now I think it is possible to distribuite this power on the whole gear surfaces so to be able to solve only a steady thermal analysis... I hope it is clear what I mean.

What do you think about?

Franco

[Zigainer]

Okay, so steady state should be fine.
But as I mentioned, you need to define a UDF for the heat flux .... sorry that I can´t hekp you with that.

[felicemastronzo]

Originally Posted by Zigainer

Okay, so steady state should be fine.
But as I mentioned, you need to define a UDF for the heat flux .... sorry that I can´t hekp you with that.

No problem and thanks for the help.
Tscüß!

[LuckyTran]

Originally Posted by felicemastronzo

Hi all,

I have a gearbox filled with oil. The motion of the internal gears heat the oil. I want to calculate the temperature distribution in the oil bath. Which BC should I set on the external boundaries (the case of the gearbox)? I do not know the temperature of the walls but only the temperature of the ambient (air). I know that the heat transfer coeff. h is depending on the delta T between the walls and the external air. If I want define that coefficient according to a known law, what should I do? Can someone help me writing an UDF?

Regards
F

air | wall | oil | gears

I just wanted to double check that a UDF is necessary.

What is your known law for heat transfer coefficient?

[felicemastronzo]

Originally Posted by LuckyTran

I just wanted to double check that a UDF is necessary.

What is your known law for heat transfer coefficient?

I have a linear trend for h like h = 0.1 DeltaT + 10. (In which DeltaT is the temperature difference between the external side of the wall and the Tinfinte of the air)

Regards
F

[LuckyTran]

Originally Posted by felicemastronzo

I have a linear trend for h like h = 0.1 DeltaT + 10. (In which DeltaT is the temperature difference between the external side of the wall and the Tinfinte of the air)

Regards
F

Where is this relation coming from? I can find little physical justification for treating HTC in this manner when it is a simple convective flow of air against a wall.

Ignoring units you have:
q" = h(DeltaT) and
h = 0.1 DeltaT + 10
so that leads to:
q" = (0.1 DeltaT^2 + 10DeltaT)

This type of heat transfer coefficient will lead to nonlinear results. I imagine the solution would also be highly unstable since the heat flux is a quadratic in temperature difference, with the potential to lead to runaway solutions and diverge. Am I overthinking this?

[felicemastronzo]

Originally Posted by LuckyTran

Where is this relation coming from? I can find little physical justification for treating HTC in this manner when it is a simple convective flow of air against a wall.

Ignoring units you have:
q" = h(DeltaT) and
h = 0.1 DeltaT + 10
so that leads to:
q" = (0.1 DeltaT^2 + 10DeltaT)

This type of heat transfer coefficient will lead to nonlinear results. I imagine the solution would also be highly unstable since the heat flux is a quadratic in temperature difference, with the potential to lead to runaway solutions and diverge.

Ok, and what is your suggestion? Which can I than describe the convection phenomena? My idea was to define h and then to define the thermal BC with the "convection option" (that means defining the heat transfer coefficient h, the free stream temperature and the eventual heat generation rate). I understand your point of view but my boss want that I try to solve the problem with this empiric equation for h. If there are not other solutions, can you help me to write the UFD to define h?

Thanks
F

[felicemastronzo]

Nobody can help me in writing the UDF?

F

[felicemastronzo]

Really nobody can help me in writing the UDF?
At least some tips....

F